Haydar:

Cyclic fatigue, of course, is the traditional way to pre-crack a fracture specimen as it involves the generation of a nominally atomically sharp crack.  In fact, ASTM, in its fracture toughness testing standard (ASTM E1820-15a Standard Test Method for Measurement of Fracture Toughness), sets limits for how high the stress-intensity range can be (relative to the fracture toughness) during such fatigue pre-cracking so that the crack doesn't become too blunt.  Usually fatigue pre-cracking is performed out of a sharp notch cycled in load-control under zero-tension or tension-tension (but low R) cyclic loading.  

However, for brittle materials, such as ceramics and  intermetallics, this can become problematic as the sample will often fracture catastrophically before a fatigue crack is induced at the notch. We have used several techniques to overcome this problem, such as cycling the sample under applied (far-field) cyclic compression loading so the crack grows due to the residual tensile stresses generated on unloading (rather than on loading).  Also starting from a very sharp notch can help to avoid catastrophic fracture in these materials; we use a radiussed micro-notching technique where a machined notch is further sharpened by "polishing" the notch root with one micro diamond paste using a new razor blade pushed back and forth under a constant load.  For very brittle materials we use all these techniques but often start from a (slant) half-chevron notch - this can be a very effective means of pre-cracking ceramics (see RH Dauskardt, DB Marshall, RO Ritchie, Cyclic fatigue-crack propagation in Mg-PSZ ceramics, J. Am. Ceram. Soc., 73 (4), April 1990, pp. 893-903).

ROR

You might find some helpful discussions in the following paper:

Carpenter RD, et al. A novel technique to generate sharp cracks in metallic/ceramic functionally graded materials by reverse 4-point bending. Scripta Materialia 2000; 43: 547-552.

Hi Haydar

With the aid of the NDT optical method such as digital shearography, one can study the crack propagation & determine the stress intensity factor of a propagating crack .refer to the attached article.

Haydar:

Cyclic fatigue, of course, is the traditional way to pre-crack a fracture specimen as it involves the generation of a nominally atomically sharp crack.  In fact, ASTM, in its fracture toughness testing standard (ASTM E1820-15a Standard Test Method for Measurement of Fracture Toughness), sets limits for how high the stress-intensity range can be (relative to the fracture toughness) during such fatigue pre-cracking so that the crack doesn't become too blunt.  Usually fatigue pre-cracking is performed out of a sharp notch cycled in load-control under zero-tension or tension-tension (but low R) cyclic loading.  

However, for brittle materials, such as ceramics and  intermetallics, this can become problematic as the sample will often fracture catastrophically before a fatigue crack is induced at the notch. We have used several techniques to overcome this problem, such as cycling the sample under applied (far-field) cyclic compression loading so the crack grows due to the residual tensile stresses generated on unloading (rather than on loading).  Also starting from a very sharp notch can help to avoid catastrophic fracture in these materials; we use a radiussed micro-notching technique where a machined notch is further sharpened by "polishing" the notch root with one micro diamond paste using a new razor blade pushed back and forth under a constant load.  For very brittle materials we use all these techniques but often start from a (slant) half-chevron notch - this can be a very effective means of pre-cracking ceramics (see RH Dauskardt, DB Marshall, RO Ritchie, Cyclic fatigue-crack propagation in Mg-PSZ ceramics, J. Am. Ceram. Soc., 73 (4), April 1990, pp. 893-903).

ROR

To follow up on the discussion, I think the notch root polishing technique mentioned by Dr. Ritchie is an interesting and controllable technique. We had a similar issue in a project regarding the analysis of fracture response in transversely-graded structures. In that project, our intention was to analyze fracture response of a Ti/TiB graded material with crack front parallel to the gradient direction, i.e. material properties varied significantly along the crack front. Flat crack fronts in transversely graded structures tend to get inclined towards the more brittle side (TiB side in our work) during the propagation stage. So, even with the most well-controlled cyclic precracking techniques, you cannot expect the formation of a uniform precrack in your material. Notch root polishing might be the best solution to such problems.

 I m so glade to read some suggestions from you, thanks a lot  for you all.  

Just to supplement what Rob has mentioned. Chevron notch, in theory, will force the crack to grow stably in any brittle materials due to its increasing crack front. However, this is not always the case; premature failure or insufficient stable crack length may lead to erroneous KIC value. One has to optimise the chevron notch geometry via systematic FE modelling. Here is a good reference to read: AM Calomino, LJ Ghosn Optimum notch configurations for the four-point bend chevron notch specimen, Int. J. of Fracture, 72, 1995, pp. 311-326